Produced by Reuters Plus for
Building an operating system for social infrastructure
Explore
The Reuters news staff had no role in the production of this content. It was created by Reuters Plus, the brand marketing studio of Reuters.
Disclaimer: The Reuters news staff had no role in the production of this content. It was created by Reuters Plus, the brand marketing studio of Reuters. To work with Reuters Plus, contact us here.
AI is no longer limited to recognizing patterns or generating information. The latest models can interpret complex situations and understand how different factors interact and evolve across virtual and real environments.
AI is also taking on a growing role in critical infrastructure, where it can help address complex challenges and enable more proactive maintenance and operations. It is already being applied to areas such as inspecting railway tracks and detecting signs of energy equipment deterioration.
“The physical world is increasingly making its way into the digital world,” says Rio Kurokawa, who is responsible for data and AI strategy at Hitachi.
Yet deploying AI in these settings poses challenges around safety and cybersecurity, where even a minor error or malfunction can have serious consequences. “Successful adoption of AI goes beyond advanced technologies. It requires a broad range of capabilities,” he says.
What’s needed is an integrated, holistic approach to applying AI reliably, responsibly, and safely, one that provides long-term improved performance, efficiency, and growth.
The next generation of social infrastructure will need more than frontier AI, says Hitachi. It’ll need the expertise that comes from decades of running the systems on which society depends. Enter HMAX.
Sense, store, think, act.
Read more about HMAX by Hitachi here
Hitachi is rising to the challenge, combining decades of experience in operating social infrastructure with advanced AI and digital technologies to deliver HMAX by Hitachi. Developed through its Lumada initiative, this suite of next-generation solutions brings the power of AI to social infrastructure.
“With HMAX, we aim to build an operating system for social infrastructure to address the challenges facing people and businesses around the world to implement AI in practice,” Kurokawa says.
HMAX reflects a growing need for an integrated approach as advances in technology drive the convergence of Information Technology (IT), Operational Technology (OT), AI and data, Kurokawa says. What decides whether that integration works with mission critical quality is expertise in all four, and the pool of companies with that is small.
Over the past 30 years, IT has become central to business decision-making, while OT has enabled companies to capture and accumulate data from products and systems through sensors. As AI models have become more capable, physical-world data can also be modeled and analyzed alongside digital information. This is the inflection point – bringing IT and OT ever closer together.
Making AI harder to misuse
A prerequisite for sustaining this operational cycle, Kurokawa says, is providing AI-ready data across IT and OT while protecting customers against unexpected failures and cyberattacks. This is particularly important in high-stakes infrastructure, where mission-critical reliability is essential.
Because AI is probabilistic by nature, hallucinations and other errors cannot be eliminated entirely. “That means we cannot rule out the possibility of a robot making a wrong move. The question, then, is how to drive that risk to zero,” he explains.
He says the company is working to minimize risks through multiple measures, such as addressing them at the design stage or incorporating safety mechanisms into the system. “Requirements vary depending on local conditions and practices, so we tailor safety-by-design in our solutions for each client.”
Many industries have traditionally developed as distinct, highly specialized fields, but advances in AI are making it possible to apply knowledge across different environments and extract greater value from accumulated data, says Kurokawa, who joined Hitachi after nearly three decades working in AI at IBM and Amazon.
The AI-driven convergence of IT and OT is giving rise to HMAX’s continuous cycle of “sense, store, think and act,” allowing people, robots, or automated guided vehicles to act, receive immediate feedback and use it to improve their decisions.
A key element of that cycle is “domain knowledge,” he says. At Hitachi, operational expertise spans a broad range of industries, from railways and manufacturing to energy and data centers. This breadth is complemented by a strong sense of unity across the organization and an unwavering commitment to quality so rigorous that, for example, railway operations are managed down to the second, as standard.
“Decades of operating at that level have created a unique advantage that is difficult to replicate. That depth of domain knowledge provides a strong foundation for orchestrating complex systems and making imitation learning effective even with relatively simple physics-based models,” says Kurokawa.
The physical world is increasingly making its way into the digital world.
The human element within social infrastructure
With AI evolving so rapidly, even experts like Kurokawa find it difficult to predict what lies ahead. “I have no idea what comes next,” he says. “Just a year ago, models such as Mythos would have been difficult to imagine. The frontier of AI is advancing so quickly that even those closest to the technology are finding it harder to anticipate what may come next.”
Looking ahead to the next decade, what seems clear is that AI will become increasingly embedded in infrastructure as organizations grapple with aging equipment and labor shortages worldwide. “By then, AI may no longer need to be explicitly called out as AI,” Kurokawa says. “Designing systems, products, and infrastructures without AI must not be an option.”
Hitachi expects demand for HMAX to grow, especially as applications for the platform expand beyond its origins in railways and into energy, industry, finance, data centers and more.
But Kurokawa emphasizes that its development must always prioritize human life and safety. “We need operational systems that can adapt AI continuously in social infrastructure,” he says. “AI may be convenient, but we cannot simply apply these technologies, deploy them and walk away. We cannot leave the responsibility for society and human lives to AI. We need to develop the technology to deliver the right AI capabilities with a reliable operational system, as we do with the Data Fabric, Cyber and AI Operations services within HMAX.”
That perspective is the essence of Hitachi’s approach with HMAX. “We will continue to evolve HMAX to realize what we refer to as the ‘harmonized society,’ one in which the environment, personal well-being and economic growth coexist,” Kurokawa says. “The social infrastructure underpinning areas such as energy, transportation, industry, cities and healthcare, needs to evolve together.”
HMAX, says Kurokawa, is the mechanism and approach that connects the data, AI, and human knowledge to continuously evolve this social infrastructure.
But the more effectively AI can understand data, the faster that cycle can run—and the greater the consequences of misuse. “That is why the challenge is no longer to make AI easier to use. We also need to make it harder to misuse,” Kurokawa says. This calls for greater resilience in AI systems, including the ability to slow or stop processes when unexpected behavior is detected, identify the source of an error, and prevent it from propagating. AI needs both an “accelerator” and a “brake,” he says.
Cyberattacks are also becoming a top agenda item for every company and are increasingly seen as a matter of “when,” not “if.” Hitachi is using advanced AI to strengthen cybersecurity, including identifying vulnerabilities in its own products during development.
Kurokawa stresses that alliances with leading AI companies are essential to keeping pace with evolving attack patterns and strengthening defenses against emerging threats, helping maintain the mission-critical reliability required for social infrastructure.
AI
Data
Information Technology (IT)
Operational Technology (OT)
The challenge is no longer to make AI easier to use. We also need to make it harder to misuse.
Explore
Rio Kurokawa is Chief Lumada Business Officer, HMAX Strategy Officer and VP, AI Strategy, at Hitachi.
The HMAX integrated approach
Read the next article in the series, How to make AI agents behave in the physical world, to discover how Hitachi’s IWIM brings orchestration, context and guardrails to industrial AI agents.
READ HERE
